Face Prediction for Cosmetic Surgeries

Model—Conditional Generative Adversarial Networks


Cosmetic surgery is a type of highly risky service and consumers usually suffer from an unsatisfying consequence after taking surgery. In this paper, we manage to predict consumers’ post-surgery appearance based on other consumers’ cases by deep learning. Specifically, we utilize generative neural networks (GANs), one type of generative model to automatically generate the post-surgery face given the pre-surgery face image. We design experiments to show that our method can increase consumers’ purchase intention. Cosmetic surgery platform or service provider (i.e., hospitals) can leverage our technique to predict consumers’ post-surgery images when they are making their purchase decision.

Keywords: Face Prediction, Generative Adversarial Networks, Cosmetic Procedure

This paper is work-in-progress